AI for Graphies

Inspiration

We have really been impressed by the Neo4j move from basic querying to formalized treatments of graphs that enable data science and machine learning. Modern applications come with new requirements. Such applications need to incorporate intelligence and learning from data. They require full context to support smart decision-making in real time. Operational applications are increasingly components of complex systems that incorporate machine learning and artificial intelligence. For actionable AI, applications need to bridge data science across operational systems and leverage context in real time. Why Graph Databases Will Power Intelligent Applications ? Today’s applications connect people, companies, assets, devices and even genomes. Mathematically speaking, these connections form a graph. Graph data structures provide a flexible, scalable foundation for these connection-oriented applications. Hence, we realized that for building up more intelligence applications, it is essential to integrate front end technology, not just only on working back end algorithms due to importance of Data, Artificial Intelligence (AI) and Graph Database technology for analytics purposes.

What it does

It is a simple generalized personalized search platform powered by AI features, graph algorithms, data analytics features in the backend. It is not limited to specific datasets, but, we focus on how graph analytic algorithms can be integrated with simple AI powered web applications or mobile app built with low code, developer productivity Neo4jGraphQL new library for fetching data from analytical queries.

How we built it

Firstly, we created in memory graph on public open movie data set for doing some data analysis using gds library in the Neo4j Sandbox backend. Here are a few steps: Step 1: Create local memory graph using grds library in neo4j browser Step 2: Run PageRank Algorithm using gds library Step 3: Write results to the database as new property of node for further utilization in the app After finishing the backend tasks, we build Neo4jGraphQL API with Apollo server for personalized search application. Everything is coded with javascript. In addition, in order to enhance the visibility of our application, we added data visualization based on the algorithm's results so that the user can see the analytical results as clearly as possible.

Challenges I ran into

As it is my very first time setup this integrated platform for formulating the web application, I have faced a lot of challenges for connecting from the front-end (online part), Neo4jGraphQL , Apollo, with Vue Apollo , to backend (offline analysis part). Furthermore, It's very challenging trying to finish everything in time for final submission by doing alone.

Accomplishments that we're proud of

I'm proud of submitting my initial project to this hackathon.

What I learned

I have learned a lot about the graph algorithms, how to use them, where and why we have to implemented such algorithms for knowledge graph applications. Now, I knew that how to build up simple web applications using Neo4jGraphQL library with Neo4j graph databases too.

What's next for AI for Graphies

In the future, It can be enhanced with AI-powered Chatbot applications using Natural Language Processing (NLP) techniques combined with Neo4j graph database. Moreover, It is need to improve advanced search, personalized search functions as well.

Built With

• gds
• javascript
• neo4j
• neo4jgraphql
• sandbox
• vue